DEA Based Benchmarking Models

Abstract

Data envelopment analysis (DEA) is a methodology for identifying the efficient or best-practice frontier of decision making units (DMUs). It is required that all DMUs under consideration be evaluated against each other in a same pool. Adding or deleting an inefficient DMU does not alter the efficient frontier and the efficiencies of the existing DMUs. The inefficiency scores change only if the efficient frontier is altered. Benchmarking is the process of comparing a DMU’s performance to the best practices formed by a set of DMUs. DEA is also called “balanced benchmarking”, because DEA considers multiple performance metrics in a single model. Under such a notion, the best practices are the benchmarks identified by DEA. However, in a more general sense, best practices do not have to be identified by DEA—they can be existing “standards”. This chapter presents two DEA-based benchmarking approaches where one set of DMUs is compared (or benchmarked) against another. One approach is called “context-dependent” DEA where a set of DMUs is evaluated against a particular evaluation context. Each evaluation context represents an efficient frontier composed by DMUs in a specific performance level. The context-dependent DEA measures the attractiveness and the progress when DMUs exhibiting poorer and better performance are chosen as the evaluation context, respectively. The other approach consists of a fixed benchmark model and a variable benchmark model where each (new) DMU is evaluated against a set of given benchmarks (standards).